A singing photo, sometimes called image singing, is a short video made from one still picture and an audio track. The AI finds the face, listens to the vocal line, and moves the lips, jaw, cheeks, and eyes so the person appears to sing the song. A good result keeps the identity, hair, clothing, and background of the original and only animates the face.
Singing photo AI works by matching sounds to mouth shapes: open vowels widen the mouth, consonants close it, and held notes keep it steady. That is why clear vocals over a light backing track tend to sync better than a dense mix, and why a front-facing portrait beats a side profile.
People use it for birthday greetings, music memes, family keepsakes, fan edits, and quick tests before producing a full music video. Because the input is a single image, expect a head-and-shoulders performance rather than full-body choreography.